PROJECT SUMMARY Understanding the genetic and environmental architecture of traits has been one of the central goals of behavioral genetics over the last fifty years. Traditional approaches using twins and families have shown that most traits, including psychiatric disorders, are highly heritable. More recently, methods that estimate heritability (h2) from single nucleotide polymorphisms (SNPs) in unrelated individuals (h2SNP) have demonstrated the importance of common variants to the genetic variation underlying complex traits. In turn, the realization that common variants are responsible for substantial trait heritability has motivated continued investment in large whole-genome datasets, which have allowed the discovery of thousands of SNPs reliably associated with complex traits. In the midst of this deluge of data, however, fundamental questions about the genetic and environmental architecture of traits remain unanswered, and new methodological approaches that leverage increasingly large whole-genome datasets are needed to answer them. In this Renewal application, we build on our previous methodological work to answer four high-level questions about the genetic and environmental architecture of complex traits. First, estimates of h2SNP for psychiatric disorders remain lower than estimates of h2 from twin and family studies. How much of this ?still missing? heritability is due to rare risk variants? Using methods developed during the previous period of our grant, we will provide the best estimates to date of the importance of rare versus common risk variants of schizophrenia, bipolar disorder, and major depression. Second, there appears to be substantial overlap between common risk alleles for psychiatric disorders such as schizophrenia and bipolar disorder. Do rare risk alleles overlap to the same degree, or do they tend to be disorder-specific? We will use extensions of our previously developed methods to help answer this question. Third, the availability of large whole-genome datasets is growing at an unprecedented rate. Can this data be leveraged to answer fundamental questions about the importance of genes and the environment, traditionally the domain of twin and family designs? We propose the development of methodological approaches that use measured genetic data among relatives that exist in large datasets to help answer old questions in new ways that bypass earlier limitations. Finally, it is crucial to understand factors that can bias estimates and lead to incorrect conclusions. We show how assortative mating and gene-by- environment interactions bias existing estimates of h2SNP, and we propose the development of models and software tools that mitigate these biases. By project's end, we anticipate having tools that open up new vistas to behavioral genetics research, allowing for a clearer understanding of the genetic and environmental architecture of psychiatric disorders and other complex traits. Doing so will help guide future analytic and investment decisions.